The Agent Skills Directory

[COMMAND_EXECUTION]: The skill employs bash to execute diagnostic commands, including curl for connectivity tests, grep and tail for log analysis, and docker commands for managing self-hosted instances. These operations are restricted to troubleshooting purposes.\n- [DATA_EXFILTRATION]: Authentication tokens (LANGFUSE_PUBLIC_KEY and LANGFUSE_SECRET_KEY) are accessed via environment variables to perform health checks against the official Langfuse API (cloud.langfuse.com). This is a documented functional requirement and uses the well-known service's own infrastructure.\n- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes data from application logs (/var/log/app/*.log) and API responses. While this represents a vulnerability surface, the risk is low given the specific troubleshooting context and the use of filtering tools.\n
Ingestion points: Application log files and external API responses.\n
Boundary markers: None present.\n
Capability inventory: Bash, Read, Write, and Edit permissions.\n
Sanitization: Content is filtered through grep and processed via python3 for data extraction.

langfuse-incident-runbook